def convert_docfile(self, input, filename, output, outputname):
result = True
try:
self.converter.convert(os.path.join(input, filename),
os.path.join(output, outputname))
except Exception as e:
logger.info(e)
result = False
return result
def wait_until_port_open(timeout=10):
count = 0
while (not utils.builtin.is_port_open(self.host, self.port)):
time.sleep(0.01)
count += 0.01
if count > timeout:
print "Can not connect unoconverter server."
logger.info("Can not connect unoconverter server.")
break
def __init__(self, fileobj, name, output_base):
super(LibreOfficeProcessor, self).__init__(fileobj, name, output_base)
if self.converter is None:
self.__class__.converter = utils.docprocessor.unoconverter.DocumentConverter(
)
self.html_path = self.output_path.html
self.docbook_path = self.output_path.docbook
self.mimetype = self.mimetype()
logger.info('Mimetype: %s' % self.mimetype)
self.result = self.convert()
def makedesktop(self):
while (True):
try:
context = self.resolver.resolve(
"uno:socket,host=localhost,port=%s;urp;StarOffice.ComponentContext"
% self.port)
self.desktop = context.ServiceManager.createInstanceWithContext(
"com.sun.star.frame.Desktop", context)
break
except Exception as e:
self.restartservice()
logger.info("%s. Restart service and rebuild desktop." % e)
if not utils.builtin.is_port_open(self.host, self.port):
raise DocumentConversionException, "failed to connect to OpenOffice.org on port %s" % self.port
def convert(self):
"""
>>> import os
>>> import shutil
>>> import utils.docprocessor.libreoffice
>>> import xml.etree.ElementTree
>>> basepath = 'core/test_output'
>>> f = open('core/test/cv_1.doc', 'r')
>>> cv1 = utils.docprocessor.libreoffice.LibreOfficeProcessor(f, 'cv_1.doc', basepath)
>>> cv1.result
True
>>> e = xml.etree.ElementTree.parse(os.path.join(
... cv1.docbook_path, cv1.name.xml)).getroot()
>>> e.findall('para')[0].text
'http://jianli.yjbys.com/'
>>> with open(os.path.join(cv1.markdown_path,
... cv1.name.md))as file:
... data = file.read()
>>> 'http://jianli.yjbys.com/' in data
True
>>> f.close()
>>> shutil.rmtree(basepath)
"""
logger.info('Convert: %s' % self.base)
if self.mimetype in [
"application/msword",
"application/vnd.openxmlformats-officedocument"
".wordprocessingml.document"
]:
if self.mimetype == "application/vnd.openxmlformats-officedocument.wordprocessingml.document":
shutil.copyfile(os.path.join(self.source_path, self.name),
os.path.join(self.docx_path, self.name.docx))
elif self.mimetype == "application/msword":
returncode = self.convert_docfile(self.source_path, self.name,
self.docx_path,
self.name.docx)
returncode = self.convert_docfile(self.docx_path, self.name.docx,
self.docbook_path, self.name.xml)
if not os.path.exists(
os.path.join(self.docbook_path, self.name.xml)):
logger.info('Not exists')
self.resultcode = 2
return False
if returncode is False:
self.resultcode = 3
return False
self.remove_note()
self.file_docbook_to_markdown()
logger.info(' '.join([self.base.base, self.name.base, 'Success']))
self.resultcode = 0
return True
else:
logger.info('Skip')
self.resultcode = 1
return False
def convert(self):
"""
>>> import os
>>> import shutil
>>> import utils.docprocessor.pandoc
>>> basepath = 'core/test_output'
>>> f = open('core/test/cv_1.docx', 'r')
>>> cv1 = utils.docprocessor.pandoc.PandocProcessor(f, 'cv_1.docx', basepath)
>>> cv1.result
True
>>> '13888888888' in cv1.markdown_stream
True
>>> f.close()
>>> shutil.rmtree(basepath)
"""
logger.info('Convert: %s' % self.base)
self.file_docx_to_markdown()
if not os.path.exists(os.path.join(self.markdown_path, self.name.md)):
self.resultcode = 2
return False
logger.info(' '.join([self.base.base, self.name.base, 'Success']))
self.resultcode = 0
return True
def convert(self, inputFile, outputFile):
retry = 0
inputUrl = self._toFileUrl(inputFile)
outputUrl = self._toFileUrl(outputFile)
loadProperties = {"Hidden": True}
self.makedesktop()
while (True and retry < 3):
try:
document = utils.timeout.thread.timeout_call(
self.desktop.loadComponentFromURL,
30,
kill_wait=1,
args=(inputUrl, "_blank", 0,
self._toProperties(loadProperties)))
document.refresh()
family = self._detectFamily(document)
outputExt = self._getFileExt(outputFile)
storeProperties = self._getStoreProperties(document, outputExt)
try:
document.storeToURL(outputUrl,
self._toProperties(storeProperties))
finally:
document.close(True)
break
except utils.timeout.exception.ExecTimeout as e:
logger.info("DocumentConverter timeout.")
logger.info("Restart service and rebuild desktop.")
self.restartservice()
self.makedesktop()
except Exception as e:
# com.sun.star.uno.RuntimeException:
# Binary URP bridge already disposed
logger.info(e)
logger.info("Restart service and rebuild desktop.")
self.restartservice()
self.makedesktop()
retry += 1